In [1]:
# load required modules
import scanpy as sc
import besca as bc
import pandas as pd
import numpy as np
import os
#import numpy as np
#from matplotlib import pyplot
#import sys
#setup document
sc.settings.verbosity = 3 # verbosity: errors (0), warnings (1), info (2), hints (3)
sc.settings.set_figure_params(dpi=80) # low dpi (dots per inch) yields small inline figures
sc.logging.print_versions()
In [2]:
# define filepath (this is the folder that contains "raw" and "analyzed")
root_path = os.getcwd()
# input: results from standard workflow
analysis_name = 'standard_workflow_besca2_0'
results_folder = os.path.join(root_path, 'analyzed', analysis_name)
input_data = os.path.join(results_folder, analysis_name + '.h5ad') # specify a .h5ad file for storing the results
# define output directories
outdir_data = results_folder
outdir_figures = os.path.join(results_folder, 'figures')
outdir_results = os.path.join(results_folder, 'results')
sc.settings.figdir = os.path.join(outdir_figures)
os.makedirs(outdir_data, exist_ok=True)
os.makedirs(outdir_figures, exist_ok=True)
os.makedirs(outdir_results, exist_ok=True)
In [3]:
#reload our data from previously written out AnnData object
adata = sc.read(input_data)
In [4]:
adata
Out[4]:
In [5]:
adata.obs.head()
Out[5]:
In [ ]:
In [6]:
%matplotlib inline
sc.settings.set_figure_params(dpi=90)
In [7]:
sc.pl.umap(adata, color=['leiden', 'Type'], legend_loc='on data', legend_fontsize=9)
In [8]:
sc.pl.umap(adata, color=['leiden', 'Subtype'], legend_loc='on data', legend_fontsize=6)
In [9]:
sc.pl.umap(adata, color=['CD4', 'CD8A'])
In [10]:
sc.pl.umap(adata, color=['Subtype'], legend_fontsize=9)
In [11]:
# Load GMT file and calculate signature scores
import pkg_resources
gmt_file_IMM=pkg_resources.resource_filename('besca', 'datasets/genesets/HumanCD45p_scseqCMs6.gmt')
adata_with_scores = adata.copy()
bc.tl.sig.combined_signature_score(adata_with_scores, gmt_file_IMM)
In [12]:
adata_with_scores
Out[12]:
In [13]:
sc.pl.umap(adata_with_scores, color=['score_HumanCD45p_scseqCMs6_Tcells_scanpy',
'score_HumanCD45p_scseqCMs6_Cellcycle_scanpy',
'score_HumanCD45p_scseqCMs6_Endothelial_scanpy'], legend_loc='on data',legend_fontsize=6)
In [14]:
mymarkers=bc.datasets.load_immune_signatures()
setName = 'HumanCD45p_scseqCMs6'
# Fract_pos was exported by BESCA in the standard worflow
f=pd.read_csv(results_folder + "/labelings/leiden/fract_pos.gct",sep="\t",skiprows=2)
df=bc.tl.sig.score_mw(f,mymarkers)
myc=np.median(df.loc[setName+'_Ubi',:]*2/3)
cNames=bc.tl.sig.make_anno(df,myc,setName,f, 0.3)
adata=bc.tl.sig.add_anno(adata,cNames, 'leiden')
In [15]:
sc.pl.umap(adata, color=['cell_names','sscell_group', 'leiden'], legend_loc='on data',legend_fontsize=6)
sc.pl.umap(adata, color=['cell_group','scell_group', 'leiden'], legend_loc='on data',legend_fontsize=6)
In [ ]:
In [ ]:
In [ ]:
In [16]:
#write down new cluster names (important order needs to be equivalent to above)
new_cluster_names = ['CD4-positive, alpha-beta memory T cell', 'CD8-positive, alpha-beta cytotoxic T cell',
'memory B cell', 'naive T cell', 'IgM or IgA plasma cell', 'regulatory T cell', 'myeloid leukocyte',
'CD8-positive, alpha-beta memory T cell', 'naive B cell', 'intestinal epithelial cell',
'IgM or IgA plasma cell', 'IgG plasma cell', 'fibroblast', 'IgM or IgA plasma cell',
'blood vessel endothelial cell', 'intestinal epithelial cell', 'group 3 innate lymphoid cell',
'group 1 innate lymphoid cell', 'memory T cell', 'proliferating T cell', 'myeloid dendritic cell',
'pericyte cell', 'neural cell', 'HEV endothelial cell', 'plasmacytoid dendritic cell']
bc.tl.annotate_cells_clustering(adata=adata, clustering_label='leiden', new_annotation_label='dblabel', new_cluster_labels=new_cluster_names)
In [17]:
#write down new cluster names (important order needs to be equivalent to above)
new_cluster_names = ['T cell', 'T cell', 'B cell', 'T cell', 'plasma cell', 'T cell', 'myeloid leukocyte', 'T cell',
'B cell', 'epithelial cell', 'plasma cell', 'plasma cell', 'fibroblast', 'plasma cell',
'endothelial cell', 'epithelial cell', 'ILC3', 'ILC1', 'T cell', 'T cell', 'cDC', 'pericyte cell',
'neural cell', 'endothelial cell', 'pDC']
bc.tl.annotate_cells_clustering(adata=adata, clustering_label='leiden', new_annotation_label='celltype', new_cluster_labels=new_cluster_names)
In [ ]:
In [18]:
#write down new cluster names (important order needs to be equivalent to above)
new_cluster_names = ['0: T cell', '1: T cell', '2: B cell', '3: T cell', '4: plasma cell', '5: T cell', '6: myeloid leukocyte',
'7: T cell', '8: B cell', '9: epithelial cell', '10: plasma cell', '11: plasma cell',
'12: fibroblast', '13: plasma cell', '14: endothelial cell', '15: epithelial cell',
'16: ILC3', '17: ILC1', '18: T cell', '19: T cell', '20: cDC', '21: pericyte cell',
'22: neural cell', '23: endothelial cell', '24: pDC']
bc.tl.annotate_cells_clustering(adata=adata, clustering_label='leiden', new_annotation_label='cluster_celltype', new_cluster_labels=new_cluster_names)
In [ ]:
In [19]:
adata = bc.st.additional_labeling(adata, 'dblabel', 'dblabel', 'Curated celltype annotation.', 'Klas Hatje', outdir_data)
In [20]:
adata = bc.st.additional_labeling(adata, 'celltype', 'celltype', 'Manual celltype annotation.', 'Klas Hatje', outdir_data)
In [21]:
adata = bc.st.additional_labeling(adata, 'Type', 'Martin2019_Type', 'Cell type annotation from authors [Martin2019].', 'Martin et al', outdir_data)
In [22]:
adata = bc.st.additional_labeling(adata, 'Subtype', 'Martin2019_Subtype', 'Cell subtype annotation from authors [Martin2019].', 'Martin et al', outdir_data)
In [ ]:
In [ ]:
adata.write(filename=os.path.join(outdir_data, 'Martin2019_processed.h5ad'), )